2X YC Alum l Co-founder + GP https://xitter.cf/t.co/bngyJFFaOu
San Francisco via NJ
Joined November 2007
- Tweets7K
- Following2K
- Followers2.5K
- Likes25.2K
Ryan Bednar retweeted
The execution velocity of @mynameisyahia and @getcontextdev is just insane. I don’t understand how it’s possible.
Ryan Bednar retweeted
5 of our portfolio cos are now solving a key piece of our internal OC tech stack
- @mastra for agent framework
- @triggerdotdev for background agents and cron
- @getcontextdev for realtime web data
- @sazabi for observability and fixing slop
- @superset_sh for ADE
and more on the way…
what a time to be an investor right now.
Ryan Bednar retweeted
Introducing overshoot, it will cut your GPU bill in half with zero quality loss, or you don’t pay.
Over the past few months, we have forward deployed into AI teams to optimize their inference stacks. They have now cumulatively saved over $5M.
Overshoot is software that bakes in our hard-earned inference learnings into a scalable solution.
overshoot.ai
Ryan Bednar retweeted
Today, we're launching Sazabi version 1.0.
The reality of shipping software as a team in the age of AI is actually crazy.
Agents merging their own PRs. Hundreds of deployments per day. A codebase that is constantly changing. Junior developers with 10x or 100x leverage. 😵💫
I spent the last year building a safety harness for teams that ship with AI. It's called @sazabi.
The premise is simple:
1. You send us logs.
2. We detect anomalies, investigate them, and send you back rich, actionable alerts about real problems in your app.
Sazabi helps you find issues before your users notice.
No monitors or noisy alerts. Just pure signal.
And today we're making it available for anyone to use.
Ryan Bednar retweeted
Coding this weekend? Need a web data API?
Your first 10,000 requests are on us
For the next 6 hours, i'm giving away 10,000 free credits on context.dev
That’s enough for any one of these:
> 10,000 scrapes, browser rendering + proxy escalation included
> 10,000 web searches
> 1,000 answers
> 1,000 brand lookups
> so much more
Comment what you’re building + DM me, I’ll add the credits myself!
Ryan Bednar retweeted
We made a bet at @Halluminate in mid-2025 that Finance would be the next big field after coding that would really take off from agent usage. Few reasons for this:
1. Finance is the world's largest knowledge work industry
2. The outputs are semi-verifiable (which has some nice properties for RL)
3. Financial reasoning generalizes to a lot of other knowledge work domains the same way SWE agent capabilities generalized to a lot of other engineering domains (ex. CAD, MLE).
Almost all my peers in banking have become increasingly AI-pilled in how they do work (similar to SWEs in Dec 2025 with Claude Code) and there's so much more progress to come soon.
Its even more exciting to see our hard work in RL Envs/Benchmarking play out in the hands of consumers.
If you'd like to contribute to building Financial Superintelligence we're hiring!
Today is the day that AI passed the "financial modeling Turing test" for me.
A "push button" build from scratch Skill that one-shotted a model on $MU that is indistinguishable (to me) from a model that a junior analyst would build from scratch. With full, impeccable adherence to all aspects of who i like to format, design & build models (took a few turns to dial it in).
Opus 5.5 is unbelievable.
Ryan Bednar retweeted
Introducing Superset for iPhone.
Remotely control agents running on your computer. Review and ship PRs from your pocket.
Ryan Bednar retweeted
To be a good founder, you have to be a very harsh judge of everything around you. Most painfully, of people.
Ryan Bednar retweeted
I gave you all the heads-up!!
Now Financial Datasets is one of the hottest companies at demo day.
1 week ago: we were his first check.
Now:
Oversubscribed. Taking strategic and 5 star investor checks.
I asked Virat how much he wanted to raise, and he said he’s not sure he even needs to raise anything.
He still has every dollar he’s ever raised in the bank.
They’re growing 31% since July.
WEEK OVER WEEK.
31% week-over-week growth and cash flow positive. Crossed 200k annualized revenue halfway through the batch and has increased growth rate as he scales.
Second derivative of revenue is positive. (My favorite metric milestone)
He can share actual revenue numbers but check out the gorgeous graph below.
Contact me if you want an intro at [email protected]
The entire internet will eventually be 99% agentic usage. He’s starting with financial data and then moving into government data, health data, insurance data - any part of the economy that has hard to access, often regulated datasets, will need agentic infrastructure to access it. It’s an enormous vision.
If you’re at Demo Day, find @virattt , he’s simultaneously the most ambitious and most thoughtful founder you’ll meet.
There is a new kind of startup that answers the call:
"Make something agents want."
which is a spin on YC's classic slogan: "Make something people want."
Financial Datasets helps agents connect to stock market data.
Yes, you a person, can sign up for Financial Datasets.
But most of their impressive growth is agents finding them, self-serve signing up, and using the product.
This is a wild timeline.
@virattt has a history of building fintech infrastructure. Proud to back him.
Read the memo: orangecollective.vc/memos/fi…
If you're interested in learning more, you can reach out for an intro. Or ask your agent to.
Ryan Bednar retweeted
i'm really enjoying the part of vibe coding that lets me get all my dumb ideas out, play with them for a bit, and then toss them. instead of just repeatedly wondering if they are good or bad ideas. wonderful for clearing room for new bad ideas, at least
Ryan Bednar retweeted
Today’s modern investment analysts are using Claude Code, Codex, and custom harnesses to research public markets.
But a harness is only as good as its system of record and the tools you give it.
Financial Datasets is a pay-as-you-go API for searching and monitoring live markets data, indexed directly from primary sources (SEC filings, earnings calls, etc).
Early devs started by building toy apps, like IPO trackers or ETF look throughs. Then Wall Street caught on.
Now it’s quickly becoming the data infrastructure for top firms, and a new financial system run by agents.
Excited to back @findatasets and @virattt!
Ryan Bednar retweeted
There is a new kind of startup that answers the call:
"Make something agents want."
which is a spin on YC's classic slogan: "Make something people want."
Financial Datasets helps agents connect to stock market data.
Yes, you a person, can sign up for Financial Datasets.
But most of their impressive growth is agents finding them, self-serve signing up, and using the product.
This is a wild timeline.
@virattt has a history of building fintech infrastructure. Proud to back him.
Read the memo: orangecollective.vc/memos/fi…
If you're interested in learning more, you can reach out for an intro. Or ask your agent to.
Ryan Bednar retweeted
Today we hit a pretty big milestone at Didit
We’re now at 64x the revenue we were doing 15 months ago. And almost 3x since YC Demo Day, just 5 months ago. All PLG
Compounding is real. It feels painfully slow… until suddenly it doesn’t 🚀
Ryan Bednar retweeted
100% agree, and this only scratches the surface.
• Satellite cities boom because who cares about an extra 20 minutes on your commute if you can sleep or work
• Traffic all but disappears if all FSD cars are on the same system and know to reroute
• DWIs are a thing of the past
• Fatal teenage driver accidents go to zero
• Shipping costs are dramatically reduced
• Mobile food trucks that prepare your food on the way to your house
• At home services explode (think haircuts, therapy, dog walking) become more practical
• Flying (especially with young families) doesn’t make sense on trips under 10 hours (just leave at night and be there in the morning)
And these are just obvious implications off the top of my head. Who knows the 100s of other 1st, 2nd, and 3rd effects that we will see.
People still aren't quite pricing in how revolutionary FSD is going to be to human life, and how transport works now.
Anecdote: I go pick up the Bangalorean at the Marin ferry to be chivalrous, but really, I don't touch the wheel at all and just put on FSD and sit in the driver's seat the entire way listening to podcasts.
There's no reason for me to be in the car. If I could, I'd just punch in the round trip into the routing, and have the car go pick her up and drive her back. It's good enough at this point to do it, just the law and custom is in the way. Humans don't need to drive anymore.
Now imagine how much of your life would change if you could send, get picked up by, or send off somewhere a comfy box with wheels (and networked screens that run AI) at any point or place in your life.
Not Uber, with the wait and get on and share a space and maybe the car stinks. Your own personal mini-lounge and storage space that you can send off or make appear anywhere.
Kids now take themselves to and back from school. Parking no longer a hassle as you can dispatch the car to find itself a place to wait (or even do robotaxi work and make you some cash). You can run anything across town, or have anything sent to you in real-time. Friend flying in? Hey, will send you a car to the airport and have it drop you off at our dinner venue. Heck, you can take a nap or do a meeting inside your pod on wheels if you're in between places.
It'll be this new mobile third place that most everyone will have, while nobody thinks about the stress or time wasted of just staring at the car in front of you and keeping it in a lane (often in slow traffic). As a side benefit of FSD, you totally tune out all the surrounding cars and aggravation associated with dealing with erratic human drivers. Someone cuts you off? Eh, you did it to AI. Someone honks...you're yelling at a bot man. Who cares? Cortisol levels just plunge when FSD is running things since you're no longer the mediating consciousness.
It probably sounds like SF scifi hokum to those who still drive, but FSD knowers know how it just utterly changes everything about moving through the world. Going back to driving feels like having to giddy-up a horse....something nobody would do if they could avoid it. Things are about to get weird.
Ryan Bednar retweeted
congrats to @PhilipJohnston and the @Starcloud_ team on the massive $250M round, and $2.3B valuation!
was lucky to catch up with Philip at YC Startup School a few weeks ago, where we chatted about how to build and raise money for ambitious ideas (spoiler alert: book a launch to space)
Philip really sets the bar on ambition and resourcefulness. goes to show that you really can just make things happen.
I'm super excited to share that we have raised $250M at a $2.3B valuation to build a LOT more AI compute capacity in orbit 🚀
Thanks so much to our supporters, including Manhattan West, who led the round, and new investors, @NVIDIA, @Cisco_Invests, @CedarCapital, Goanna Capital, and @Standard_Cap, as well as continued investment from existing backers @Benchmark, @EQTVentures, @Soma_Capital, @NFX, @SevenSevenSix, and others.
Next week, @Starcloud_ will move into a new 100,000 square ft facility that will enable us to ramp up our production rate to 100 satellites/week.
This is just the beginning! @EzraFeilden, @AdiOltean🏃🏻➡️🏃🏻➡️🏃🏻➡️
Ryan Bednar retweeted
What if it wasn't the reactor that's making nuclear power plants expensive, but another overlooked part?
Apollo Atomics gave me a peek at how they've shrunk the steam generator, which they say will help nuclear beat natural gas on price. techcrunch.com/2026/08/20/ap…
Ryan Bednar retweeted
Fewer than five people on Earth have built two open-source JavaScript framework companies. @calcsam is one of them.
The book that made him known is now one of the go-to texts for building agents. You’ve seen it: Principles of Building AI Agents.
But the part that’s important to his story is the low point before @Mastra. When @GatsbyJS faded, the acquisition went sideways. They went off to build sales AI and failed for months, then came back to open-source dev tools and instantly felt like they were back on track.
His word for it: the "unknown knowns." The things you know so well you forget you know them. They'd forgotten they were world-class at the exact thing they walked away from.
I sat down with him at @ycombinator, where Mastra went through the Winter 2025 batch.
We got into:
•Why "that's interesting" is the most dangerous thing a user can tell you
•Why planning in quarters is dead, it's actions-per-minute now
•Why it's okay to quit your idea
•Writing a print book in an era that moves too fast to print
•The loneliness nobody talks about: teams stopped pair programming because everyone's driving their own agent
His rule, from a YC partner: generosity breeds luck.
𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒
(00:00) Intro — Y Combinator
(01:30) What Mastra Is
(03:00) Content Wants to Be Frequent
(05:00) Shipping a Tier-One Feature Every Day
(07:00) The $13M Seed and $22M Series A
(09:00) Turn-Based vs. Real-Time Strategy
(12:00) Why Experience Beats Fresh Eyes
(15:00) "That's Interesting" Means They're Not Interested
(18:00) It's Okay to Quit Your Idea
(21:00) The Book That Started as a Joke
(25:00) From Journalism to Open Source
(28:00) The Netlify Acquisition, and What It Cost
(32:00) The Unknown Knowns
(36:00) The YC Winter '25 Takeoff Story
(40:00) Remote Work Is Underrated in the Age of AI
(44:00) The New Loneliness of Building With Agents
(47:00) What He Still Can't Figure Out About Agents